institutional investors views and preferences on climate ... · institutional investors’ views...
TRANSCRIPT
1
Institutional Investors’ Views and Preferences on Climate Risk Disclosure
Emirhan Ilhan Philipp Krueger Zacharias Sautner Laura T. Starks
April 2019
Abstract
We survey institutional investors on firm-level climate risk disclosure. Many investors believe that climate risk reporting is as important as traditional financial reporting. However, current quantitative and qualitative disclosures on climate risks are regarded as being insufficient and imprecise. Many investors believe that climate reporting should be mandatory and more standardized. Disclosure is seen as more important for assessing physical climate risks, and less important for regulatory risks. Investors that believe that current climate-related disclosures are deficient perceive stronger underpricing of climate risks in equity markets.
_________________________
Emirhan Ilhan is at the Frankfurt School of Finance & Management ([email protected]). Philipp Krueger is at the University of Geneva and the Swiss Finance Institute ([email protected]). Zacharias Sautner is at the Frankfurt School of Finance & Management ([email protected]). Laura T. Starks is at the McCombs School of Business, University of Texas at Austin ([email protected]).
2
1. Introduction
The efficiency of financial markets relies on timely and accurate information about the
exposure of firms to important risks. Many firms increasingly face risks related to climate change,
originating from natural disasters or regulation to combat a global rise in temperature (Litterman
2016, Krueger, Sautner, and Starks 2019). High-quality information on firms’ exposures to climate risks
has therefore become relevant for investors to make informed investment decisions, and for markets
to correctly price the risks and opportunities related to climate change.
However, many regulators and investors argue that climate risk disclosure is currently
insufficient. For example, Mark Carney, Governor of the Bank of England, demanded that more needs
to be done “to develop consistent, comparable, reliable and clear disclosure around the carbon
intensity of different assets” (Carney 2015). In a similar spirit, Anne Stausboll, former CEO of CalPERS,
argued that “consistent and comparable corporate disclosure of material climate issues is critical [and
that] investors require better climate disclosure“ (Stausboll 2014). More recently, Yngve Slyngstad,
CEO of Norges Bank Investment Management, echoed these concerns by commenting on the difficulty
of obtaining climate risk-related data: “The only surprise […] is how hard it is to get the data […] I think
it will take years to get good data from the majority of companies we are invested in.”1
To address potential shortcomings in current climate disclosures, regulators, governments,
and NGOs started to take actions aimed at improving firm-level reporting on climate risks. For
instance, the Financial Stability Board initiated in 2015 the Task Force on Climate-related Financial
Disclosures (TCFD), with the objective to develop voluntary climate-related financial risk disclosures.
In a similar spirit, CDP collects climate-related information by means of a voluntary questionnaire on
behalf of investors representing over $87tr in assets under management. On top of these largely
voluntary initiatives, U.K. securities law mandates since 2013 the disclosure of carbon emissions for
1 See “Norway wealth fund builds tool to analyze climate risk to portfolio,” Reuters Market News, October 31, 2018.
3
quoted companies (see Krueger 2015; Jouvenot and Krueger 2019), and U.S. lawmakers are currently
developing mandatory climate-risk disclosure.2 In France, institutional investors are required since
2016 by law to report the carbon footprints of their investment portfolios.
While these initiatives suggest that investors increasingly require climate-related information
for their decision making, little systematic evidence exists on institutional investors’ beliefs with
respect to such disclosures. In this paper, we directly survey institutional investors about their views
and preferences with respect to climate-related disclosures. We are able to link the survey responses
to data on investor characteristics which we also collect by means of the survey. The main benefit of
our survey is that it allows us to shed light on important climate-related investor perspectives and
actions that cannot be studied using archival data. Our respondent group consists of important
decision makers at some of the world’s largest investors. About one-third of the respondents works
at the executive level in their institutions, and 11% work for institutions with more than $100bn in
assets under management. Surveys are increasingly used in the finance literature, enabling better
understandings of such topics as corporate financing (Graham and Harvey 2001), investor activism
(McCahery, Sautner, and Starks 2016), investor relations (Karolyi and Liao 2017), or ESG investing
(Amel-Zadeh and Serafeim 2018).
In the first part of the analysis, we survey investors’ views regarding the importance of firm-
level climate reporting, the shortcomings of available reporting, and the need for mandatory and
standardized reporting. This analysis enables us to identify areas on which firms or regulators should
focus to satisfy the disclosure demands of investors. In the second part, we provide some first
evidence on the relation between investor beliefs regarding the quality of current climate disclosures
and perceived climate risk mispricing in equity markets. In the third part, we address important recent
2 Recently, the Climate Risk Disclosure Act of 2018 was introduced in the U.S. Senate, with the objective to introduce mandatory climate disclosure. If accepted, the bill would require firms to disclose greenhouse-gas emissions, risk-management strategies to address climate change, and discussion about how climate change affects valuations in different climate scenarios.
4
developments in climate disclosure, namely disclosure-motivated engagement with portfolio firms
and the disclosure of the investors’ own portfolio-level carbon footprint.
Reporting on non-financial information can have benefits and costs (Christensen, Hail, and
Leuz 2019), so it is theoretically ambiguous whether investors attribute value to climate risk disclosure
by firms. For example, while such reporting may increase stock liquidity, reduce a firm’s costs of
capital, and make the pricing of risks more efficient, it may also allow competitors to infer proprietary
information about a firm’s future strategy (e.g., if future CO2 emission targets are disclosed).
Our institutional investors share a strong general belief that climate disclosure is important.
In fact, 51% of respondents believe that climate risk reporting is as important as traditional financial
reporting, and almost one-third even considers it to be more important. Only 22% of respondents
regard climate reporting as less (or much less) important compared to financial reporting. Climate
disclosure is perceived as more important among those investors who also believe more strongly that
climate risks are important, and among those who expect larger global temperature increases due to
climate change.
Climate change can affect the portfolio firms through three channels. Physical climate risks
arise because of adverse effects of changes in the physical climate (e.g., temperatures increases).
Technological climate risks originate from climate-related innovations that disrupt traditional
producers (e.g., when electric car manufacturers may displace traditional manufacturers), and
regulatory risks result from costs associated with changes in policies or regulations to combat climate
change (e.g., carbon taxes).
Climate disclosure is deemed most important by those investors that worry strongly about the
financial consequences of these three types of climate risks for their portfolio firms. In terms of their
relative importance, concerns about physical climate risks matter the most for the perceived
importance of climate reporting, while regulatory risks matter the least. An implication of this finding
5
is that disclosure is likely to be most valuable when it enables investors to better evaluate the physical
climate risks of firms. One reason may be that physical risks tend to be more firm and location specific,
thus requiring precise information about a firm’s exposure to evaluate them correctly. Regulatory
risks, on the other hand, are usually more firm-independent and regulator-dependent, and
information on such risks may be easier to obtain from sources outside of the firm.
The vast majority of our respondents believe that current quantitative and qualitative
disclosures on climate risks are uninformative and imprecise. Many investors also share the view that
climate risk reporting should be mandatory and standardized, as it is currently the case with financial
reporting. Investors that worry more about the financially effects of climate risks also agree more
strongly that there should be more and better disclosure on climate risks. Our respondents also
believe that investors should put pressure on firms to disclose more information about their climate
risks. This widespread view echoes recent investor initiatives at Exxon Mobil and Occidental
Petroleum, where a group of institutions submitted shareholder proposal calling for these firms to
share more information on their climate policies.3
Next, we build on recent theoretical work that predicts a link between climate mispricing and
disclosure (Daniel, Litterman, and Wagner 2017). Investors’ views on the availability and quality of
current climate reporting are strongly related to the perceived underpricing of climate risks in equity
markets (i.e., climate-related overvaluation of firms). Notably, respondents who believe that current
reporting is lacking also observe more mispricing in current equity valuations. An important
consequence of this finding is that better disclosure may contribute to the more efficient pricing of
climate risks. This implication is consistent with the view expressed by Michael R. Bloomberg, Chair of
3 See “Exxon Shareholders Pressure Company on Climate Risks,“ The Wall Street Journal, May 31, 2017 and “Occidental Shareholders Vote for Climate Proposal,” The Wall Street Journal, May 31, 2017.
6
the TCFD, that “increasing transparency makes markets more efficient, and economies more stable
and resilient.”4
In terms of embracing current developments in climate disclosure, the majority of our
investors engage or plan to engage portfolio firms to report according to the recommendations of the
TCFD.5 Further, our respondents indicate support for the French approach to require also reporting
on the carbon footprints of the investment portfolios: 60% of respondents either currently disclose or
plan to disclose the CO2 footprints of their portfolios. This result indicates support for ongoing policy
efforts at the European Union level that head in the same direction.
In terms of generalizability of our findings, we should note that our respondent group is likely
biased toward investors with a relatively high awareness of climate risks. The reason is that such
investors are probably more disposed to participate in a climate risk survey. In addition, some of our
responses were obtained at ESG conferences. Nevertheless, understanding the views and preferences
of such investors is particularly important, because they are more likely to shape future disclosure
policies through industry initiates (e.g., CDP, CERES, or UN-PRI) or lobbying with regulators.
Our paper contributes to a relatively scarce academic literature on climate-related disclosure.
Solomon et al. (2011) interview institutional investors who reveal that they use private channels of
discourse with their portfolio firms to compensate for the inadequacies of public climate reporting.
Matsumura, Prakash, Vera-Muñoz (2014) find that markets discount firms that do not disclose carbon
emissions, and Plumlee et al. (2015) find a positive association between disclosure quality and firm
value. Ilhan, Sautner, and Vilkov (2019) show that information about carbon risks (if disclosed) are
used by investors, as firms with larger carbon emissions exhibit higher tail risk. Matsumura, Prakash,
4 See https://www.fsb-tcfd.org/. 5 These recommendations include disclosing climate-related risks and opportunities and their impact on firms’ businesses; how firms governance structures deal with these risks and opportunities; how firms identify, assess, and manage climate risks; and which metrics and targets firms use to assess and manage carbon emissions.
7
and Vera-Muñoz (2018) analyze voluntary 10-K climate risk disclosures and find that disclosers have
lower costs of equity than non-disclosers.
Krueger (2015) examines the valuation effects of the introduction of mandatory greenhouse-
gas (GHG) disclosures in the U.K., and shows beneficial valuation effect resulting from the regulation.
More recently, Jouvenot and Krueger (2019) examine the real effects of the introduction of mandatory
GHG reporting in the U.K. They document strong reductions in carbon emissions for U.K. firms relative
to control firms from other jurisdictions. Focusing on the oil and gas industry, Eccles and Krzus (2018)
examine to what extent firms disclose information in line with the TCFD recommendations.
Our main contribution to this literature is detailing institutional investors’ views and actions
on climate-related disclosures. More generally, we also contribute to the literature on non-financial
(or sustainability) reporting, of which climate risks are an important component. Christensen, Hail,
and Leuz (2019) review the current literature on sustainability reporting.
2. Methodology and Survey Design
2.1 Survey Development and Delivery
We used both an online and a paper version of the survey that we distributed through four
delivery channels, yielding a total of 439 responses.6 First, we personally distributed the paper version
of the survey at four institutional investor conferences: The Sustainable Investment Conference in
Frankfurt on November 9, 2017; the ICGN Paris Event on December 6-7, 2017; the Asset Management
with Climate Risk Conference at Cass Business School in London on January 23, 2018; and the ICPM
Conference in Toronto on June 10-12, 2018. We obtained a total of 72 responses from these four
conferences.
6 The survey instrument is provided in the Internet Appendix. We used an iterative process for developing the survey. Details of this process are provided in Krueger, Sautner, and Starks (2019). The original survey also contained questions on climate risk management and shareholder engagement, which are covered in Krueger, Sautner, and Starks (2019).
8
Second, we distributed the online version of the survey to 1,018 individuals in senior functions
at institutional investors. We identified these individuals using the help of a survey service provider
that manages a global panel of more than 5m registered professionals. The panel contains detailed
data on these individuals’ job titles and their age to identify relevant subsamples. The provider invited
these individuals in March 2018 to participate in the survey and we obtained 410 initial responses. We
then excluded 90 participants that took less than five minutes to complete the survey and participants
for which basic checks yielded logical inconsistencies in the responses (Meade and Craig 2012). This
process left us with 320 responses of good quality. These respondents took on average 15 minutes to
complete the survey. The service provider had several mechanisms in place to ensure the authenticity
of the participants.
Third, in April 2018, we emailed invitations to participate in the survey to a list of institutional
investors that cooperate with a major asset owner through CERES and IIGCC on climate risk topics. We
obtained 28 responses through this channel. Fourth, we sent invitations to participate in the online
survey to personal contacts at different institutional investors, yielding 19 additional responses.
We are confident that in the vast majority of cases we have only one observation per
institution. The reason is that, for 87% of the observations, key identifying characteristics do not
coincide.7 In the remaining cases we cannot exclude the possibility that respondents work for the
same institution. However, the responses are sufficiently different among these respondents to
discount that possibility with some degree of assurance. We provide a discussion of potential survey
response bias in Krueger, Sautner, and Starks (2019).
2.2 Respondent Characteristics
Table 1 provides an overview of the characteristics for our respondent groups. The largest
numbers of respondents are fund or portfolio managers (21%), followed by executive or managing
7 These characteristics are location, assets under management, institutional investor type, investor horizon, ESG share (+/–10% variation in the variable), equity share (+/–10%), and passive share (+/–10%).
9
directors (18%). About one-third works at the executive level in their institutions, as CIO (11%), CEO
(10%), or in related functions (10%). Most respondents work for asset managers (23%) and banks
(22%), followed by pension funds (17%), insurance companies (15%), and mutual funds (8%). Our
sample includes 23% of respondents that work for institutions with assets between $20bn and $50bn,
16% with assets between $50bn and $100bn, and 11%, with assets of more than $100bn.
Only 5% of respondents’ institutions typically hold investments for less than six months, 38%
have medium holding periods (six months to two years), 38% have long holding periods (two years to
five years), and the remaining 18% typically hold investments for more than five years. Our
respondent’ institutions are headquartered around the world: 32% are located in the U.S., 17% in the
U.K. and Ireland, 12% in Canada, and 11% in Germany, among others. The average portfolio share of
our respondents’ institutions that incorporates ESG aspects is 41%, they invest on average 47% in
equities (43% in fixed income), and 38% of their assets are on average passively invested.
Table 2 reports summary statistics of other survey variables that we study in the subsequent
sections. These variables will be discussed in detail below.
3. Evidence on Climate Risk Disclosure
3.1 Importance of Climate Risk Disclosure
Reporting on non-financial information through CSR or climate risk reporting can have benefits
and costs to investors. On the one hand, non-financial disclosure can increase stock liquidity by
alleviating adverse selection among investors (Verrecchia 2001). Reporting on non-financial
information can also lower the cost of capital of portfolio firms (Plumlee et al. 2015, Matsumura,
Prakash, and Vera-Muñoz 2018) and may allow for a better pricing and hedging of climate risks. On
the other hand, non-financial disclosure can be costly to portfolio firms if it reveals proprietary
information to competitors (Ellis, Fee, and Thomas 2012). Such costs are less relevant for high level or
aggregated disclosures, but they can be substantial for detailed disclosures. For example, if a firm
10
discloses detailed carbon reduction targets, this may allow competitors to infer a firm’s future product
market strategy. In light of these benefits and costs, it is theoretically ambiguous what importance
investors attribute to reporting on climate risks.
To evaluate this theoretical ambiguity, we asked respondents to indicate how important they
consider reporting on climate risks relative to reporting on financial information (Question B1). The
corresponding responses are reported in Figure 1. We find that 51% of respondents believe that
climate risk reporting is as important as financial reporting, and almost one-third even considers it to
be more important. Interestingly, only 22% of respondents regard climate risk reporting as less or
much less important compared to traditional financial reporting. Overall, these figures imply that
there is a broad belief that climate reporting is important.
Next, we perform in Table 3 regressions that shed light on how investor beliefs about the
relative importance of climate reporting vary in the cross-section. Building on the responses displayed
in Figure 1, the dependent variable in the table is an integer-valued measure that varies between one
(climate risk reporting is much less important) and five (climate risk reporting is much more
important). Across the difference columns in the table, we try to evaluate whether climate reporting
is perceived as more important if investors attribute also a higher importance to climate change and
its financial impact on portfolio firms.
As a starting point, Columns (1) examines whether climate reporting is perceived as more
important among investors that attribute a higher general importance to climate risks when investing
in firms. We measure the importance of climate risks using Climate risk ranking, which is the ranking
an investor attaches to climate risk considerations when making investment decisions. To construct
this variable, we asked the respondents to indicate the importance of climate risks relative to other
investment risks (Question A1).8 The resulting rank variable ranges between one (climate risks are
8 These other risks included traditional financial risks (e.g., earnings risk), operational risks, governance risks, social risks, and other environmental risks.
11
most important) and six (climate risks are least important). Summary statistics on this variable are
reported in Table 2, and Krueger, Sautner, and Starks (2019) discuss the variable in more detail.
Next to this variable, we control in this and subsequent regressions for investor horizon,
investor size, an investor’s ESG and passive shares, whether an investor is an independent institution
(Ferreira and Matos 2008), and the social and environmental norms in an investor’s country (Dyck et
al. 2019). We further include fixed effects for the respondents’ positions in their institutions and for
the distribution channels.
Examining the coefficient estimate on Climate risk ranking in Column (1), we find that those
investors who rank climate risks higher also believe that climate reporting is more important. The
effects are economically meaningful. Compared to the median investor—who regards climate
reporting to be just as important as financial reporting—an investor who ranks climate risks one
standard deviation higher deems climate reporting 16% ((-0.3*-1.6)/3) more important.
Next, Column (2) examines whether the respondents’ climate change expectations help
explain the importance of climate reporting. To elicit expectations, we used the 2°C target of the Paris
Climate Accord as an anchor, and then asked the respondents about their own global temperature
expectations by the of this century (Question E1).9 Responses could vary between one (no expectation
of a temperature rise) and five (more than 3°C expected). Across all respondent, only 3% do not expect
any temperature increase by 2100, 16% expect an increase by up to 1°C, and 30% by up to 2°C. Four
in ten respondents expect a temperature rise that exceeds the Paris 2°C target, with 12% expecting
an increase of more than 3°C.
Personal climate expectations seem to be highly relevant for explaining the perceived
importance of climate reporting. Column (2) indicates that investors who expect a higher temperature
rise also consider climate reporting to be more important. A standard deviation increase in the
9 Under the 2015 Paris Climate Accord, 195 countries agreed to take significant measures to keep the global temperature rise under 2°C by the end of this century.
12
expected temperature rise, which corresponds to moving up one notch in the possible response
category, corresponds to a 10% (0.34/3) higher value for the importance attached to climate
disclosures.
Climate change can affect the value of portfolio firms through three channels. Physical
climate risks can affect firms because of severe weather events, draughts, or rises in sea levels.
Regulatory risks encompass costs that result from regulations aimed at reducing the negative impacts
of climate change.10 Technological risks relate to climate-related technological disruption that may
adversely affect portfolio firms. In our survey, we asked the investors to rate the financial materiality
of each of these risks and then construct three variables to measure their financial effects (Regulatory
climate risk, Physical climate risk, and Technological climate risk). Each of the variables can vary
between one (not at all important) and five (very important).
Columns (3) to (5) examine whether the importance of climate disclosure varies across
investors based on how financially material they evaluate each of these three risks. Across all columns,
we find that investors who deem climate risks more financially material also attach greater importance
to climate reporting. This finding is consistent with Amel-Zadeh and Serafeim (2018) who find that
investment professionals consider ESG information whenever they consider it financially material to
investment performance. Most interestingly, the three coefficient estimates differ substantially in
magnitude. The estimate in Column (4) on Physical climate risk is almost 2.5 times as large as the
estimate on Regulatory climate risk in Column (3). In a similar but less pronounced way, the estimate
on Technological climate risks in Column (5) is substantially larger than the estimate on Regulatory
climate risk in Column (3).11 These differences suggest that disclosure is seen as most important when
it comes to physical climate risks, followed by technological and then regulatory risks.
10 Examples for such regulation include a carbon tax such as the one proposed in the Washington Initiative 1631. 11 As the distribution of the three risk variables is almost identical (see Table 2), we can directly compare the coefficient estimates to evaluate their relative importance. The coefficient on Regulatory climate risk is
13
The strong role of physical risks in explaining the importance of climate disclosure may be
because such risks tend to be more firm and location specific, requiring relatively precise information
about a firm’s exposure to evaluate them. Regulatory climate risks, on the other hand, are more firm-
independent and regulator-dependent, and information on such risks may be easier to obtain since
firms in the same industry and country face similar regulatory risks (e.g., information could also be
obtained from competitors). In addition, Krueger, Sautner, and Starks (2019) show that a relatively
large fraction of investors believe that regulatory climate risks have already started materializing,
while physical and technological risks are expected to materialize over longer horizons only. The more
immediate character of regulatory risks might therefore imply that disclosure about them is less
important than information about (potentially more distant) technological and physical risks.
Turning to the control variables, we find that investors with more assets under management
believe that climate reporting is more important, which is plausible as such investors tend to be
universal owners and stand to lose more from climate risks. Perhaps surprisingly, medium- and long-
term investors do not differ from short-term investors in their perceptions of the importance of
climate reporting. More ESG oriented investors generally regard climate disclosure as being more
important, which is not surprising since the investment mandate of such investors is partially based
on environmental aspects. Investors based in more climate-conscious countries generally regard
climate disclosure as being more important, consistent with Dyck et al. (2019).
3.2 Evaluation of Current Disclosure Practice
Investors attribute great importance to climate risk reporting by portfolio firms, but such
reporting is still in its infancy and largely voluntary and unstandardized. 12 It is therefore important to
better understand investors’ views on the informativeness of current qualitative and quantitative
statistically different from those on Physical climate risk and Technological climate risk, while the latter two do not differ statistically. 12 To the best of our knowledge, the only country that entertains a mandatory and prescriptive climate disclosure regime is the U.K., which introduced this with the The Companies Act 2006, Regulations 2013). For an analysis of this regulation see Krueger (2015).
14
disclosures on climate risks (Question B3). Qualitative disclosure may include narratives of how
climate change affects business models and how climate risks are governed in firms, while quantitative
disclosures can contain information on carbon emissions or emission reduction targets. To assess
investors’ views on these disclosure, respondents were asked to indicate their agreement with a set
of statements on these topics on a scale of one (“strongly disagree”) through five (“strongly agree”).
Table 4 shows that there exists a widespread view that current quantitative and qualitative
disclosures are not sufficiently informative and also imprecise. Specifically, many of our respondents
strongly agree that management discussions on climate risk (20.8% strongly agree) as well as
quantitative information on these risks (19.4% strongly agree) are not sufficiently precise. These
numbers suggest that that the current voluntary reporting regime does not enable fully informed
investment decisions, at least for firms with large exposures to climate risks. This could be one reason
why climate risks are considered difficult to price in equity markets, an issue we address in more detail
below.
The responses to the previous two questions suggest that many firms currently do not
consider the net benefits of reporting on climate risks to be sufficiently high, as they would otherwise
reveal such information voluntarily. At the same time, investors seem to value such information, which
raises the question whether mandatory and standardized reporting on climate risks is needed. In
general, the economic rationale for mandatory disclosure regulation on climate risks requires the
existence of externalities and/or market-wide cost savings that regulations can mitigate (Shleifer
2005). A firm’s contribution to climate change can be viewed as such an externality. Standardization
of climate reporting could make it easier and less costly for investors to acquire and interpret climate
risk information and it could facilitate cross-firm and cross-industry benchmarking. A mandatory
disclosure regime could also provide commitment and credibility for firms’ climate disclosure,
especially if the standards are specific and well enforced (Christensen, Hail, and Leuz 2019).
15
Indeed, Table 4 documents that many investors believe that standardized and mandatory
reporting on climate risk would be valuable (27% strongly agree). Similarly, there is a widespread view
among our respondents that there should be more standardization in climate-related financial
disclosure across markets (27%). However, one challenge for changing the current reporting
environment seems to be that standardized disclosure tools and guidelines are not yet widely available
(21.3%), at least according to our respondents.
Overall, our respondents’ views are consistent with recent initiatives that increase
transparency on climate topic. For example, in June 2017, TCFD released its recommendations on
climate-related financial disclosures, which centers on the role of climate risks for a firm’s governance,
strategy, and its risk management, and how climate risks are reflected in metrics and targets. Using
this comprehensive approach, the TCFD recommendations go well beyond simply disclosing carbon
emissions.
Though still voluntary, it is likely that the TCFD recommendations will eventually constitute
the basis for mandatory and standardized climate disclosure in many countries. At the same time, it
will most likely take considerable amounts of time until such regulations are eventually introduced in
a large set of countries. In the meantime, investors themselves may take initiatives to improve their
access to climate risk data. In fact, many of our respondents hold the strong belief that investors
should put pressure on firms to disclose more on their climate risks (28% strongly agree), echoing the
recent investor initiatives at Exxon Mobil or Occidental Petroleum discussed earlier. This view of our
respondents is also in line with the increasing role that climate topics played in the most recent proxy
seasons (see Ceres, 2018).
Table 5 provides regressions explaining investors’ views on current disclosure practices.
Building on the responses in Table 4, the dependent variables in the table equal one if a respondent
indicated “strong agreement” with a statement on the current disclosure practices, and zero
otherwise. Our main explanatory variable that might drive investors’ views about current disclosures
16
is Climate risk materiality, which measures how respondents evaluate the financial effects of climate
risks (Question A2). This variable averages the responses to the three questions about the materiality
of regulatory, physical, and technological climate risks, and it can range between one (“not at all
important”) and five (“very important”).
Columns (1) and (2) indicate that investors who believe that climate risks are more material
also think more strongly that current information on climate risks is imprecise and uninformative.
These investors also believe more strongly that there should be more standardization and mandatory
requirements in climate disclosures (see Columns (3) and (4)). The results in Column (7) further suggest
that investors who believe strongly that climate risks matter also strongly believe that investors should
demand better disclosure from portfolio firms.
Turning to our control variables, we find that investors whose portfolios are more subject to
ESG analysis believe more strongly that quantitative information on climate risks is imprecise. They
also tend to agree more strongly that management discussions and disclosure forms are lacking in
quality and informativeness. In addition, investors with more assets under ESG principles also believe
more strongly that tools and guidelines for standardized disclosure are currently not available.
3.3 Climate Risk Disclosure and Climate Risk Mispricing
Recent research suggests that climate risks may be mispriced. On the empirical side, Hong, Li,
and Xu (2019) document that markets underreact to climate risks because of poor disclosure,
suggesting that value exists in improving disclosure. On the theoretical side, Daniel, Litterman, and
Wagner (2017) develop a model in which uncertainty about the effect of CO2 emissions on global
temperature (and on eventual damages) is gradually resolved over time. Their model suggests a high
carbon price today that is expected to decline over time as uncertainty about climate risks is resolved.
One mechanism through which these uncertainties disappear is via climate risk disclosures. As firms
evaluate the risks climate change poses on their business models and make their assessments public,
17
equity prices should converge towards their fair valuations through the harmonization and
comparability benefits of disclosures.
To contribute to this important literature, we first asked the investors whether they believe
that current equity valuations in different sectors of the economy correctly reflect the risks and
opportunities related to climate change (Question D1).13 Responses for each sector could vary
between plus two (valuations much too high) and minus two (valuations much too low). Figure 2
reports the mean responses across sectors, showing that overvaluations are highest in the oil and
automotive sector. (Krueger, Sautner, and Starks (2019) provide more discussion on how
overvaluation varies across sectors.)
We use these data to create two dependent variables. For each respondent, Climate risk
underpricing averages all positive mispricing scores across sectors (score of one or two), to capture
the extent to which a respondent believes that there is climate-related overvaluation. Relatedly,
Climate risk mispricing averages for each respondent the absolute values of the mispricing scores
across all sectors, to capture nondirectional mispricing. Table 2 shows that the average respondent
believes that equity valuations in the average sector do not fully reflect the risks from climate change,
as the mean value of Climate risk underpricing exceeds zero.
We next estimate whether investors’ views on climate disclosure help partially explain any
perceived climate risk mispricing. We focus on three independent variables that are informative about
our respondents’ views on climate reporting. The first variable measures whether a respondent
“strongly agrees” that investors should demand more disclosure from portfolio firms about their
exposure to climate risks (Demand more disclosure). The other two variables capture perceptions
13 We allowed for over- and undervaluation across different sectors as some sectors may be overvalued (e.g., the oil or coal sectors), while other sectors may be undervalued (e.g., the battery producers, water utilities).
18
about the quality of available climate information, both in terms of hard (Quantitative information
imprecise) and soft information (Management discussion imprecise).
Table 6 reports the corresponding results. The estimate in Column (1) indicates that
respondents who more strongly agree that investors should demand disclosure on climate risks also
see stronger overvaluations. In terms of magnitudes, climate risk-related overvaluations are almost
35% (=0.2/0.57) higher, relative to the mean score of 0.57, among respondents who strongly agree
that investors should demand more disclose on climate risks. In Columns (2), we also find more
perceived overvaluation among investors who believe that the available quantitative information
about climate risks are imprecise. We find similar results in Column (3) for investors who think that
management discussions on climate risk are not sufficiently precise. Taken together, this suggests that
a lack of hard and soft information on climate risks contributes to the perception of climate risk
underpricing in equity markets.
Columns (4) to (6) confirm these results using the measure that captures both directions of
mispricing. The fact that we also find similar effects for this alternative variable suggests that better
climate disclosure is useful in alleviating both directions of climate risk mispricing, i.e., the
underpricing and overpricing of climate risks. Overall, the evidence in Table 6 indicates that the
demand for climate-related information, and beliefs about the quality of climate-related disclosure,
are associated with mispricing in equity markets, at least as perceived by our respondents. An
important implication of our evidence is that better disclosure may contribute to a more efficient
pricing of climate risks.
3.4 Recent Trends in Climate risk Disclosure
We used two questions to evaluate recent trends in climate risk disclosure. In the first
question, we evaluated the respondents’ views related to a new investor practice championed by a
French law, which requires since 2016 carbon reporting on the portfolios of institutional (Article
19
173).14 Our respondents indicate support for this approach, which is considered to be one of the most
ambitious climate risk regulations in the world: 60% stated in response to Question B2 that they
already disclose or plan to disclose the carbon footprint of their own investment portfolios (Figure 3).
This result also speaks to ongoing policy efforts at the European Union level. Under Article 7 of the
European Commission’s action plan on sustainable finance, it is discussed to amend EU Directive
2016/2341 (IORP 2-Pensions), which would require increased disclosures by institutional investors
relating to sustainability risks.
In the second question, we asked whether the investors engage or plan to engage portfolio
firms to report in accordance with the TCFD recommendations (Question E5). This is a highly relevant
topic given that several major investors recently announced that this will be a prime area for their
shareholder engagement (Blackrock 2017). Figure 4 shows that this approach is shared widely, as 59%
of investors plan to engage firms on this topic. Interestingly, a quarter of our survey participants
responded with “Do not know”, which could indicate that some institutional investors are still
unaware of the TCFD recommendations.
In Table 7, we run regressions to understand which investor characteristics explain investors’
behavior in terms of action according to these recent developments. We use two dependent variables:
Carbon footprint equals one if an investor discloses (or plans to disclose) the carbon footprint of the
own portfolio, and zero if not. TCFD equals one if an investor engages (or plans to engage) portfolio
firms to report according to the recommendations of the TCFD.
Column (1) indicates that investors who believe that climate risks are more financially material
are also more likely to disclose the carbon footprint of their portfolios. Investors with more assets
under management, and investors whose portfolios have higher ESG shares, are also more likely to
disclose their own carbon footprints. These findings are plausible as investors who believe in the
14 See “France Gets Climate Risks Disclosures from Invest Firms,” Wall Street Journal, December 7, 2017. The law also requires investors to report on how they identify and manage climate risks.
20
financial materiality of climate risks, and those that are more ESG oriented, have stronger incentives
to make the carbon footprint of their portfolios publicly available. To the extent that calculating and
disclosing the carbon footprints of portfolios is costly, it is also unsurprising that bigger institutions are
more likely to do so (perhaps because they have more financial resources). Larger investors might also
face more scrutiny by stakeholders on these issues, making them more likely to initiate actions. The
latter argument is consistent with Krueger, Sautner, and Starks (2019), who find that reputational
concerns are one of the most important drivers for institutional investors to incorporate climate risk
into the investment process. Contrary to our expectation, medium- and long-term investors have a
lower propensity to disclose/plan to disclose the carbon footprint of their portfolios compared to
short-term investors.
Column (2) suggest that investors with bigger ESG shares are more likely to engage their
portfolio firms to report according the TCFD recommendations. We also find that investors in
countries where environmental issues are seen to be more important are more likely to engage their
portfolio firms over climate risk disclosure, which is consistent with Dyck et al. (2019). We do not find
that investors differ in their likelihood of engaging portfolio firms along dimensions of climate risk
materiality, investment horizons, or assets under management. This suggests that the adoption of
these recommendations is widespread among a variety of institutional investors.
We close our analysis by examining whether and how investor adoption of recent
developments in climate disclosure relate to investor demand about more and better reporting on
climate risks. We test several predictions. First, to more accurately disclose the carbon footprint of
their portfolios, investors may want more high-quality reporting on climate risks. Such investors may
also believe more strongly that investors should demand disclosures from their portfolio firms.
Second, investors who engage or plan to engage firms to report according to the TCFD
recommendations may see a stronger need for better quality, and more standardized, disclosures.
21
Table 8 reports results from regressions which use as dependent variables our measures of
investor beliefs about the current disclosure trends. We focus on three aspects, namely the quantity
of available information, the quality of information, and standardization and mandatory reporting
requirements (see Question B3). To capture the adoption of current trends in climate reporting, we
use the explanatory variables Carbon footprint and TCFD.
The estimates indicate that investors who disclose (or plan to disclose) the carbon footprint
of their portfolios demand more disclosure about climate risks, and they also believe that such firms’
disclosure should be in a more standardized and mandatory way. Investors planning or already
disclosing their portfolio footprint also believe that management discussions about climate risks are
currently imprecise.
Given that calculating portfolio footprints is primarily about the use of quantitative
information, we find surprisingly little evidence that investors who (plan to) disclose their footprints
perceive current quantitative information as imprecise. Hence, it seems that firm-level quantitative
information relevant for investors’ carbon footprint calculations is not overly insufficient. This is
possibly the result of initiatives such as CDP, which already collects emissions data by means of a
survey. Nevertheless, there is a strong view among investors who (plan to) disclose their carbon
footprints that climate disclosures should be more standardized and mandatory. This is consistent
with the idea that, to disclose a portfolio carbon footprint, standardized issuer-level information is
required and that such data is often not available for all firms in the same format. This interpretation
is echoing the concerns by Yngve Slyngstad of Norges Bank Investment Management (see
Introduction). Indeed, Jouvenot and Krueger (2019) show that mandatory and prescriptive carbon
requirements dramatically increase the availability of carbon data at the issuer level.
We further find that investors that plan to engage firms to report according to the TCFD
recommendations see a stronger need for more disclosure, and they also believe more strongly that
disclosure should be standardized and mandatory. On the other hand, such investors do not seem to
22
think that current quantitative information or management discussions are imprecise. Thus, investors
seem to be interested in TCFD recommendations as it provides them with one way to impose some
structure on climate risk reporting towards their portfolio firms (rather than as a way to obtain more
precise information).
4. Conclusion
We use a global survey of institutional investors to examine firm-level climate risk disclosure.
A large majority of our survey respondents believes that climate risk reporting by portfolio firms is
important. In fact, many respondents consider it as important, or even more important, than reporting
on traditional financial risks. At the same time, a widespread view exists that climate risk disclosure
needs improvement, in terms of the availability and quality of hard and soft climate-related
information. Many investors further believe that firm-level reporting should be more standardized
and mandatory.
In cross-sectional tests, we find that investors who think that climate risks are more financially
material also deem climate disclosure to be more important. In a similar spirit, investors who expect
larger global temperature increases by the end of the century also believe that climate disclosure is
more important. Our analysis also suggests that firm-level disclosure seems more important for
assessing physical and technological climate risks, and less so for regulatory risks.
The views on the availability and quality of climate-related disclosures are associated with
investor-level perceptions of climate risk mispricing in the equity market. Respondents who believe
that investors should require firms to report on climate risks, and investors who regard both
quantitative and qualitative climate information to be insufficient, perceive more mispricing in current
equity valuations. Finally, the majority of our respondents plans to engage portfolio firms to report
23
according to the TCFD recommendations. A majority of investors also discloses or plans to disclose the
carbon footprint of the own investment portfolio.
Our analysis is important because through our survey we are able to shed light on many
important investor perspectives and actions that cannot be studied using archival data. This enables
us to contribute to the emerging literature on climate finance and, more generally, to the literature
on non-financial disclosure.
24
References
Amel-Zadeh, Amir, and George Serafeim, 2018, Why and how investors use ESG information: Evidence from a global survey, Financial Analysts Journal 74, 1-17.
Blackrock, 2017, How BlackRock Investment Stewardship engages on climate risk, Blackrock Market
Commentary March 2017.
Carney, Mark, 2015, Breaking the tragedy of the horizon-climate change and financial stability, Speech 29 September, Bank of England, Lloyds of London.
Ceres, 2018, The Hidden Story of Climate Proposals in the 2018 Proxy Season, available at http://bit.ly/2L4eJpv.
Christensen, Hans B., Luzi Hail, L., and Christian Leuz, 2019, Economic analysis of widespread adoption of CSR and sustainability reporting standards, Working Paper, Chicago Booth School of Business.
Daniel, Kent D., Robert B. Litterman, and Gernot Wagner, 2017, Applying asset pricing theory to
calibrate the price of climate risk, Working Paper, Columbia Business School.
Dyck, Alexander, Karl Lins, Lukas Roth, and Hannes Wagner, 2019, Do institutional investors drive corporate social responsibility? International evidence, Journal of Financial Economics 131, 693-714.
Eccles, Robert G., Michael P. Krzus, and George Serafeim, 2011, Market interest in nonfinancial information, Journal of Applied Corporate Finance 23, 113–127.
Ellis, Jesse A., Edward C. Fee, and Shawn Thomas, 2012, Proprietary costs and the disclosure of
information about customers, Journal of Accounting Research 50, 658-727.
Ferreira, Miguel A., and Pedro Matos, 2008, The color of investors’ money: The role of institutional
investors around the world, Journal of Financial Economics 88, 495-533.
Graham, John R., and Campbell R. Harvey, 2001, The theory and practice of corporate finance:
Evidence from the field, Journal of Financial Economics 60, 187-243.
Hong, Harrison G., Frank Weikai Li, and Jiangmin Xu, 2019, Climate risks and market efficiency, Journal
of Econometrics, 208, 265-281.
Ilhan, Emirhan, Zacharias Sautner, and Grigory Vilkov, 2019, Carbon tail risk, Working Paper, Frankfurt School of Finance & Management.
Jouvenot, Valentin and Philipp Krueger, 2019, Reduction in Corporate Greenhouse Gas Emissions Under Prescriptive Disclosure Requirements, Working Paper, University of Geneva.
Karolyi, George Andrew, and Rose C. Liao, 2017, The economic consequences of investor relations: A
global perspective, Working Paper, Cornell University.
Krueger, Philipp, 2015, Climate change and firm valuation: Evidence from a quasi-natural experiment, Working Paper, Swiss Finance Institute.
Krueger, Philipp, Zacharias Sautner, and Laura T. Starks, 2019, The importance of climate risks for institutional investors, Working Paper.
25
Litterman, Bob, 2016, Climate Risk: Tail Risk and the Price of Carbon Emissions---Answers to the Risk Management Puzzle. Hoboken, NJ: John Wiley & Sons.
Matsumura, Ella M., Rachna Prakash, and Sandra C. Vera-Munoz, 2014, Firm-value effects of carbon emissions and carbon disclosures, The Accounting Review 89, 695-724.
Matsumura, Ella M., Rachna Prakash, and Sandra C. Vera-Munoz, 2018, Capital market expectations of risk materiality and the credibility of managers’ risk disclosure decisions, Working Paper.
McCahery, Joseph, Zacharias Sautner, and Laura T. Starks, 2016, Behind the scenes: The corporate
governance preferences of institutional investors, Journal of Finance 71, 2905-2932.
Meade, Adam, and Bart Craig, 2012, Identifying careless responses in survey data, Psychological Methods 17, 37-55. 17, 37-55.
Plumlee, Marlene, Darrell Brown, Rachel M. Hayes, and R. Scott Marshall, 2015, Voluntary environmental disclosure quality and firm value: Further evidence, Journal of Accounting and Public Policy 34, 336-361.
Shleifer, Andrei, 2005, Understanding regulation, European Financial Management 11, 439-451.
Solomon, Jill F., Aris Solomon, Simon D. Norton, Nathan L. Joseph, 2011, Private climate change reporting: an emerging discourse of risk and opportunity?, Accounting, Auditing & Accountability Journal 24, 1119-1148.
Stausboll, Anne, 2014, Analyzing climate risk: Why mandatory reporting matters, Institutional Investors, Retrieved from http://www.institutionalinvestor.com/blogarticle/3371660/analyzing-climate-risk-why-mandatory-reporting-matters/investors-pensions.html.
Verrecchia, Robert E., Essays on disclosure, Journal of Accounting and Economics 32, 97-180.
26
Data Appendix
Variable Definition Survey Question
Importance climate risk disclosure
This variable measures how important investors consider reporting by portfolio firms on climate risks compared to reporting on financial information. The variable ranges between one and five, with one indicating that climate risk reporting is “much less importance” and five indicating that it is “much more important”.
Question B1
Climate risk ranking This variable is the outcome of a ranking of the importance of climate risks relative to other more standard investment risks. The variable ranges from one (if climate risks are considered the most important risk) to six (if they are considered the least important risk).
Question A1
Regulatory climate risk This variable measures the financial materiality of regulatory climate risk. The variable can range between one (not at all important) and five (very important).
Question A2
Physical climate risk This variable measures the financial materiality of physical climate risk. The variable can range between one (not at all important) and five (very important).
Question A2
Technological climate risk
This variable measures the financial materiality of technological climate risk. The variable can range between one (not at all important) and five (very important).
Question A2
Temperature rise expectation
Temperature rise expectation measures investors’ expectations about what the global temperature rise will be by the end of the 21st century. This variable can vary between one (no expectation of a temperature rise) and five (more than 3°C expected).
Question E1
Climate risk materiality
This variable averages the responses to three questions about the financial materiality of regulatory, physical, and technological climate risk. Each of these three variables can range between one (not at all important) and five (very important).
Question A2
Management discussions imprecise
This variable takes the value of one if a respondent strongly agrees that management discussions on climate risk are not sufficiently precise, and zero otherwise.
Quant. information imprecise
This variable takes the value of one if a respondent strongly agrees that firm-level quantitative information on climate risk is not sufficiently precise, and zero otherwise.
Stand. and mandatory reporting necessary
This variable takes the value of one if a respondent strongly agrees that standardized and mandatory reporting on climate risk is necessary, and zero otherwise.
More standardization This variable takes the value of one if a respondent strongly agrees that there should be more standardization across markets in climate-related financial disclosure, and zero otherwise.
Disclosure tools not available
This variable takes the value of one if a respondent strongly agrees that standardized disclosure tools and guidelines are currently not available, and zero otherwise.
Disclosure forms not informative
This variable takes the value of one if a respondent strongly agrees that mandatory disclosure forms are not sufficiently informative regarding climate risk, and zero otherwise.
Demand more disclosure
This variable takes the value of one if a respondent strongly agrees that investors should demand that portfolio firms disclose their exposure to climate risk, and zero otherwise.
Climate risk underpricing
This variable averages positive mispricing scores (negative scores are set to zero). The variable ranges between plus two (strong average overvaluation) and zero (no average overvaluation).
Question D1
Climate risk mispricing This variable averages the absolute values of all mispricing scores and can take the values of zero, one, and two.
Question D1
27
Carbon footprint This variable takes the value of one if a respondent discloses or plans to disclose the overall carbon footprint of their portfolio, and zero otherwise.
Question B2
TCFD This variable takes the value of one if a respondent engages or plans to engage portfolio companies to report according to the recommendations of the Task Force on Climate related Financial Disclosures, and zero otherwise.
Question E5
Medium horizon This variable takes the value one if the indicated typical holding period of an institutional investor is between six months and two years, and zero otherwise.
Question G2
Long horizon This variable takes the value one if the indicated holding period of an institutional investor is above two years, and zero otherwise.
Question G2
Assets under management
This variable indicates the size of an institutional investor and takes the values of one (assets under management less than $1bn); two (between $1bn and $20bn); three (between $20bn and $50bn); four (between $50bn and $100bn); and five (more than $100bn).
Question G6
ESG share This variable is the percentage of the institution’s portfolio that incorporates ESG issues
Question G5
Passive share This variable is the percentage of the institution’s portfolio that is passively managed.
Question G4
Independent institution
This variable takes the value one if an institutional investor is considered to be an independent institution, and zero otherwise. As in Ferreira and Matos (2008) and Dyck et al. (2019), independent institutions are more likely to collect information, have fewer potential business relationships with the corporations they invest in, and therefore are anticipated to be more involved in monitoring management. We classify mutual funds, asset managers, hedge funds, private equity funds, and public pension funds as independent institutions.
Question G1
HQ Country Norms
This variable captures the importance of environmental issues in the country in which an institutional investor is headquartered. The data are from Dyck et al. (2019), who construct the variable based on the Environmental Performance Index obtained from the Yale Center for Environmental Law (Yale University) and the Center for International Earth Science Information Network (Columbia University) for 2004. Larger numbers reflect a stronger common belief in the importance of environmental issues.
Question G7
Figure 1: Disclosure of Climate Risks
This figure is illustrates how important investors consider reporting by portfolio firms on climate risks compared to reporting on financial information (Question B1). We ask survey participants how important they consider reporting by portfolio firms on climate risk compared to reporting on financial information. The data are based on the responses of 439 individuals that participated in our survey.
4%
18%
51%
18%
10%
0%
10%
20%
30%
40%
50%
60%
Much lessimportant
Less important Equallyimportant
More important Much moreimportant
Figure 2: Climate Risk Mispricing
This figure reports invertors’ beliefs about whether current equity valuations in different sectors correctly reflect the risks and opportunities related to climate change (Question D1). Responses for each sector could vary between plus two (valuations much too high) and minus two (valuations much too low). The figure reports the mean response scores per sector. The data are based on the responses of 439 individuals that participated in our survey.
0.20
0.25
0.30
0.35
0.40
0.45
0.50
0.55
Mea
n r
esp
on
se s
core
Figure 3: Carbon-Footprint Disclosure by Investors
This figure reports whether which fraction of investors discloses or plans to disclose the overall carbon footprint of their portfolios (Question B2). The data are based on the responses of 439 individuals that participated in our survey.
24%
60%
16%
0%
10%
20%
30%
40%
50%
60%
70%
No Yes Do not know
Figure 4: Engagement of Portfolio Firms to Report According to TCFD Recommendations
This figure reports information about whether the investors engage or plan to engage their portfolio firms to report according to the recommendations of the Task Force on Climate-related Financial Disclosures (TCFD) (Question E5). The data are based on the responses of 439 individuals that participated in our survey.
17%
59%
24%
0%
10%
20%
30%
40%
50%
60%
70%
No Yes Do not know
Table 1: Survey Respondent Characteristics
This table provides summary statistics on the characteristics of the 439 individuals that participated in our survey. As not all respondents provided information on all characteristics, the number of observations used in the different parts of the table can fall below 439. We report data on the distribution channel, position of the responding individuals (Question G8), type of institution (Question G1), institution size (Question G6), investment horizon (Question G2), geographic distribution (Question G7), the ESG share (Question G5), the equity and fixed-income share (Question G3), and the passive share (Question G4).
Distribution channels (N=439) Percentage Investor horizon (N=432) Percentage
Panel 73 Short (less than 6 months) 5 Conferences 16 Medium (6 months to 2 years) 38 Asset owner 6 Long (2 years to 5 years) 38 Personal 4 Very long (more than 5 years) 18
Respondent Position (N=428) Percentage Region (N=429) Percentage
Fund/Portfolio manager 21 United States 32 Executive/Managing director 18 United Kingdom 17 Investment analyst/strategist 16 Canada 12 CIO 11 Germany 11 CEO 10 Italy 7 CFO/COO/Chairman/Other executive 10 Spain 5 ESG/RI specialist 10 The Netherlands 4 Other 2 France 3
Institutional investor type (N=439) Percentage Others (<3%) 9
Asset manager 23 Investment structure Mean
Bank 22 ESG share (N=415) 40.6% Pension fund 17 Equity share (N=400) 47.0% Insurance company 15 Fixed-income share (N=402) 43.1% Mutual fund 8 Passive share (N=419) 38.2%
Other institution 15
Assets under management (N=430) Percentage
Less than $1bn 19 Between $1bn and $20bn 32 Between $20bn and $50bn 23 Between $50bn and $100bn 16 More than $100bn 11
Table 2: Descriptive Statistics
This table provides summary statistics of the main survey-based variables that we use in the regression analysis. The data are based on the responses of 439 individuals that participated in our survey. As not all respondents provided answers to all questions, the number of observations used in the different parts of the table can fall below 439. Detailed variable definitions are in the Data Appendix.
Variable Mean STD Median Obs. Survey
Question
Importance climate risk disclosure 3.1 0.9 3.0 416 B1
Climate risk ranking 4.0 1.6 4.0 386 A1
Temperature rise expectation 3.3 1.0 3.0 342 E1
Regulatory climate risk 3.8 1.0 4.0 393 A2
Physical climate risk 3.5 1.1 4.0 393 A2
Technological climate risk 3.8 1.0 4.0 393 A2
Climate risk materiality 3.7 0.8 3.7 393 A2
Demand more disclosure 0.3 0.4 0.0 413 B3
Quantitative information imprecise 0.2 0.4 0.0 413 B3
Management discussions imprecise 0.2 0.4 0.0 413 B3
Standardized and mandatory reporting necessary 0.3 0.4 0.0 413 B3
Disclosure forms not informative 0.2 0.4 0.0 411 B3
More standardization 0.3 0.4 0.0 412 B3
Tools not available 0.2 0.4 0.0 413 B3
Climate risk overpricing 0.6 0.4 0.5 357 D1
Climate risk mispricing 0.8 0.4 0.7 357 D1
Carbon footprint 0.7 0.5 1.0 327 B2
TCFD 0.8 0.4 1.0 304 E5
Medium horizon 0.8 0.4 1.0 432 G2
Long horizon 0.2 0.4 0.0 432 G2
Assets under management 2.7 1.3 2.0 430 G6
ESG share 0.4 0.3 0.3 415 G5
Passive share 0.4 0.2 0.4 419 G4
Independent Institution 0.5 0.5 1.0 439 G1
HQ Country Norms 0.6 0.1 0.6 425 G7
Table 3: Importance of Climate Risk Disclosure
This table reports ordered logit regressions explaining the perceived importance of disclosure (relative to financial disclosure). The dependent variable, Importance climate risk disclosure, measures how important investors consider reporting by portfolio firms on climate risks compared to reporting on financial information. The variable ranges between one and five, with one indicating that climate risk reporting is “much less importance” and five indicating that it is “much more important” (see Figure 1). We use the following independent variables: Climate risk ranking is the outcome of a ranking of the importance of climate risks relative to other risks. The variable ranges from one (most important risk) to six (least important risk). Temperature rise expectation measures investors’ expectations about what the global temperature rise will be by the end of this century (Question E1). Regulatory, physical, and technological climate risk measure the financial materiality of regulatory climate risk, physical climate risk and technological climate risk (Question A2). All three variables can range between one (not at all important) and five (very important). This variable can vary between one (no expectation of a temperature rise) and five (more than 3°C expected). We additionally control for: Medium horizon; Long horizon; Assets under management; ESG share; Passive share; Independent institution; HQ Country Norms. Detailed variable definitions are in the Data Appendix. t-statistics (reported in parentheses) are based on standard errors that are clustered at the investor-country level. ***, **, * indicate significance levels of 1%, 5%, and 10%, respectively.
Importance climate risk disclosure
(1) (2) (3) (4) (5)
Climate risk ranking -0.30***
(-4.37)
Temperature rise expectation 0.34***
(2.93)
Regulatory climate risk 0.30***
(4.05)
Physical climate risk 0.71***
(6.58)
Technological climate risk 0.53***
(6.57)
Medium horizon -0.22 0.08 -0.11 -0.2 -0.2
(-0.52) (0.16) (-0.21) (-0.33) (-0.48) Long horizon -0.1 -0.03 -0.14 -0.37 -0.22
(-0.20) (-0.05) (-0.23) (-0.50) (-0.36) Assets under management 0.21*** 0.25** 0.23** 0.18* 0.23**
(2.70) (2.43) (2.41) (1.93) (2.52) ESG share (x100) 0.83 0.98** 0.88* 0.66 0.7
(1.54) (2.37) (1.80) (1.56) (1.56) Passive share (x100) -0.01 -0.22 0.07 -0.11 0.01
(-0.03) (-0.49) (0.18) (-0.24) (0.03) Independent Institution -0.05 0.01 -0.15 -0.07 -0.19
(-0.17) (0.04) (-0.62) (-0.29) (-0.81) HQ Country Norms 1.28 1.46 1.59 2.48*** 2.12** (1.34) (1.00) (1.50) (3.14) (2.16)
Respondent Position FE Yes Yes Yes Yes Yes Distribution Channel FE Yes Yes Yes Yes Yes
Obs. 361 326 370 370 370 Pseudo R2 0.06 0.05 0.05 0.09 0.07
Table 4: Evaluations of Current Practice of Climate Risk Disclosure
This table reports survey responses to questions on different aspects of the current climate risk disclosure practice (Question B3). Respondents were asked to indicate their agreement with different statements on a scale of one (“strongly disagree”) through five (“strongly agree”). Column (1) presents the percentage of respondents indicating strong agreement with a statement. Column (2) reports the mean score, where higher values correspond to stronger agreement. Column (3) reports the number of respondents. Column (4) reports the results of a t‐test of the null hypothesis that each mean score is equal to 3 (neither agree nor disagree). *** indicates statistical significance at the 1% levels. Column (5) reports the results of a t‐test of the null hypothesis that the mean score for a given reason is equal to the mean score for each of the other reasons, where significant differences at the 10% level are reported.
% with 5 (“strongly
agree”) score Mean score N
H0: Mean Score
= 3
Significant differences
in Mean Score vs.
Rows
Views on climate risk disclosure (1) (2) (3) (4) (5)
(1) Management discussions on climate risk are not sufficiently precise
20.8% 3.78 413 *** 1-4, 7
(2) Firm-level quantitative information on climate risk is not sufficiently precise
19.4% 3.77 413 *** 1-4
(3) Standardized and mandatory reporting on climate risk is necessary
26.9% 3.91 413 *** 4-7
(4) There should be more standardization across markets in climate-related financial disclosure
27.4% 3.92 412 *** 4-7
(5) Standardized disclosure tools and guidelines are currently not available
21.3% 3.64 413 *** 1-3, 5-6
(6) Mandatory disclosure forms are not sufficiently informative regarding climate risk
17.8% 3.70 411 *** 1-3, 5
(7) Investors should demand that portfolio firms disclose their exposure to climate risk
27.6% 3.90 413 *** 4-7
Table 5: Explaining Views on Current Practice of Climate Risk Disclosure
This table reports logit regressions explaining investors’ views on the current climate risk disclosure practice (Question B3). Respondents could indicate their agreement with different statements on a scale of one (“strongly agree”) through five (“strongly disagree”). The dependent variables are dummy variables that equal one if a respondent indicated strong agreement with a statement on the current disclosure practice, and zero otherwise. We use the following independent variables: Climate risk materiality (larger numbers reflect greater perceived importance); Medium horizon; Long horizon; Assets under management; ESG share; Passive share; Independent institution; HQ Country Norms. Detailed variable definitions are in the Data Appendix. t-statistics (reported in parentheses) are based on standard errors that are clustered at the investor-country level. ***, **, * indicate significance levels of 1%, 5%, and 10%, respectively.
Management
discussions imprecise
Quant. information imprecise
Stand. and mandatory reporting necessary
More standardization
Disclosure tools not available
Disclosure forms not
informative
Demand more
disclosure
(1) (2) (3) (4) (5) (6) (7)
Climate risk materiality 0.65** 0.47** 0.55*** 0.81*** 0.59* 0.51*** 0.90*** (2.34) (2.28) (3.73) (3.98) (1.90) (2.60) (3.88)
Medium horizon 0.83 -0.37 0.64 0.14 0.58 -0.58 -0.65 (0.95) (-0.46) (0.79) (0.15) (0.84) (-0.83) (-0.72)
Long horizon 1.08 0.12 0.75 0.18 0.57 -0.69 -0.99 (1.26) (0.16) (1.00) (0.34) (1.04) (-0.88) (-1.20)
Assets under management 0.11 0.02 -0.13 -0.08 0.23* 0.15*** -0.04 (1.20) (0.16) (-1.44) (-0.85) (1.81) (3.46) (-0.94)
ESG share (x100) 1.67*** 1.01** 1.47*** 0.45 1.19** 1.14*** 0.32 (3.73) (2.14) (3.15) (0.96) (2.21) (2.78) (0.52)
Passive share (x100) -0.46 0.41 0.67 -0.00 0.63 -0.65 0.80 (-0.71) (1.04) (1.01) (-0.01) (1.32) (-0.99) (1.26)
Independent institution 0.29 0.11 -0.30 -0.42 0.53** 0.33 -0.14 (0.51) (0.35) (-1.55) (-1.25) (2.20) (1.23) (-1.04)
HQ Country Norms 2.49 -1.38 -0.16 -0.80 0.55 0.35 0.06 (1.38) (-0.82) (-0.12) (-0.53) (0.32) (0.20) (0.04)
Respondent Position FE Yes Yes Yes Yes Yes Yes Yes Distribution Channel FE Yes Yes Yes Yes Yes Yes Yes
Obs. 369 369 369 369 369 369 369 Pseudo R2 0.13 0.09 0.11 0.11 0.10 0.11 0.15
Table 6: Climate Risk Disclosure and Climate Risk Mispricing
This table reports OLS regressions explaining perceptions of climate risk mispricing. We use two dependent variables to capture the respondents’ views on the mispricing of climate risks (Question D1). Climate risk underpricing averages positive mispricing scores (negative scores are set to zero). The variable ranges between plus two (strong average overvaluation) and zero (no average overvaluation). Climate risk mispricing averages the absolute values of all mispricing scores across all industries. We use the following independent variables: Demand more disclosure equals one if a respondent indicated strong agreement to the statement that investors should demand that portfolio firms disclose their exposure to climate risk, and zero otherwise (Question B3). Quantitative information imprecise equals one if a respondent indicated strong agreement to the statement that firm-level quantitative information on climate risk is not sufficiently precise, and zero otherwise (Question B3). Quantitative information imprecise equals one if a respondent indicated strong agreement to the statement that management discussions on climate risk are not sufficiently precise, and zero otherwise (Question B3). We additionally control for: Climate risk materiality (bigger numbers reflect greater perceived importance); Medium horizon; Long horizon; Assets under management; ESG share; Passive share; Independent institution; HQ Country Norms. Detailed variable definitions are in the Data Appendix. t-statistics (reported in parentheses) are based on standard errors that are clustered at the investor-country level. ***, **, * indicate significance levels of 1%, 5%, and 10%, respectively.
Climate risk underpricing Climate risk mispricing
Average across all sectors Average across all sectors
(1) (2) (3) (4) (5) (6)
Demand more disclosure 0.20*** 0.16***
(4.29) (3.28)
Quantitative information imprecise 0.24**
0.24*** (2.84)
(4.79)
Management discussions imprecise 0.22*** 0.19***
(3.53) (3.98)
Climate risk materiality -0.01 -0.01 -0.01 0.02 0.02 0.02
(-0.43) (-0.14) (-0.16) (0.73) (0.64) (0.70) Medium horizon -0.03 -0.04 -0.07 0.01 0.01 -0.02
(-0.27) (-0.30) (-0.54) (0.12) (0.06) (-0.15) Long horizon -0.00 -0.04 -0.06 0.03 -0.01 -0.02
(-0.04) (-0.39) (-0.54) (0.28) (-0.05) (-0.18) Assets under management 0.03 0.03 0.03 0.00 0.00 -0.00
(1.59) (1.34) (1.36) (0.06) (0.00) (-0.17) ESG share (x100) 0.29*** 0.28*** 0.26** 0.19*** 0.18** 0.16*
(3.60) (3.09) (2.48) (3.01) (2.52) (2.03) Passive share (x100) -0.02 -0.00 0.01 -0.05 -0.03 -0.02
(-0.21) (-0.05) (0.08) (-0.49) (-0.34) (-0.18) Independent institution -0.03 -0.04 -0.05 -0.03 -0.04 -0.04
(-0.47) (-0.58) (-0.82) (-0.49) (-0.59) (-0.82) HQ Country Norms -0.20 -0.16* -0.29* -0.30* -0.26 -0.37**
(-1.63) (-1.82) (-2.10) (-2.11) (-1.68) (-2.25)
Respondent Position FE Yes Yes Yes Yes Yes Yes Distribution Channel FE Yes Yes Yes Yes Yes Yes
Obs. 343 343 343 343 343 343 Adjusted R2 0.06 0.07 0.06 0.03 0.05 0.03
Table 7: Recent Trends in Climate Risk Disclosure
This table reports logit regressions explaining recent trends in climate risk disclosure. We use two dependent variables. Carbon footprint equals one if a respondent discloses or plans to disclose the overall carbon footprint of their portfolio, and zero if the investor does not (Question B2). TCFD equals one if a respondent engages or plans to engage portfolio firms to report according to the recommendations of the Task Force on Climate related Financial Disclosures (Question E5), and zero otherwise. We use the following independent variables: Climate risk materiality; Medium horizon; Long horizon; Assets under management; ESG share; Passive share; Independent institution; HQ Country Norms (larger numbers reflect a stronger belief in the importance of environmental issues in an institutions’ country). Detailed variable definitions are in the Data Appendix. t-statistics (reported in parentheses) are based on standard errors that are clustered at the investor-country level. ***, **, * indicate significance levels of 1%, 5%, and 10%, respectively.
Carbon footprint TCFD
(1) (2)
Climate risk materiality 0.31*** 0.23*
(3.71) (1.69) Medium horizon -0.72* -0.21
(-1.87) (-0.44) Long horizon -1.03* -0.24
(-1.79) (-0.61) Assets under management 0.28* 0.04
(1.73) (0.23) ESG share (x100) 1.07*** 2.36***
(2.81) (2.95) Passive share (x100) 1.00 0.23
(0.95) (0.46) Independent Institution 0.29 -0.08
(1.15) (-0.35) HQ Country Norms 0.62 6.75***
(0.36) (4.81)
Respondent Position FE Yes Yes Distribution Channel FE Yes Yes
Obs. 306 275 Pseudo R2 0.07 0.11
Table 8: Recent Disclosure Trends and Assessment of Climate Risk Disclosure
This table reports logit regressions explaining investors’ views on the current climate risk disclosure practice (Question B3). We use four dependent variables that reflect the respondents’ agreement to different statements on a scale of one (“strongly agree”) through five (“strongly disagree”). The four dependent variables are dummy variables that equal one if a respondent indicated strong agreement with a statement on the current disclosure practice, and zero otherwise. We use the following independent variables: Carbon footprint is a dummy variable equal to one if a respondent discloses or plans to disclose the overall carbon footprint of their portfolio and zero otherwise (Question B2). TCFD is a dummy variable equal to one if a respondent engages or plans to engage portfolio firms to report according to the recommendations of the Task Force on Climate related Financial Disclosures (Question E5). The following independent variables are also included: Medium horizon; Long horizon; Assets under management; ESG share; Passive share; Independent institution; HQ Country Norms. Detailed variable definitions are in the Data Appendix. t-statistics (reported in parentheses) are based on standard errors that are clustered at the investor-country level. ***, **, * indicate significance levels of 1%, 5%, and 10%, respectively.
Demand more disclosure
Quantitative information
imprecise
Management discussions imprecise
Standardized and mandatory reporting
necessary
(1) (2) (3) (4) (5) (6) (7) (8)
Carbon footprint 1.25** 0.40 0.44* 0.53**
(2.37) (1.08) (1.88) (2.01)
TCFD 0.87** 0.51 0.82 0.96***
(2.22) (1.27) (1.62) (3.10)
Medium horizon -0.38 -0.39 -0.60 -0.19 0.73 0.79 0.68 0.77
(-0.48) (-0.49) (-0.73) (-0.23) (0.80) (0.90) (0.91) (0.98) Long horizon -0.13 -0.39 0.14 0.33 1.52* 1.49* 1.28 1.25
(-0.17) (-0.48) (0.13) (0.34) (1.67) (1.85) (1.64) (1.40) Assets under management -0.11 -0.06 -0.01 -0.08 0.09 0.01 -0.20** -0.23**
(-1.20) (-0.69) (-0.08) (-0.75) (1.20) (0.14) (-2.13) (-2.47) ESG share (x100) 0.48 0.89* 0.80* 0.77 1.77*** 1.87*** 1.34*** 1.23***
(0.87) (1.89) (1.94) (1.55) (5.29) (2.84) (3.22) (2.91) Passive share (x100) 0.71 0.50 1.04** 0.30 -0.04 -0.20 0.99* 0.37
(0.74) (0.66) (2.27) (0.79) (-0.09) (-0.31) (1.90) (0.59) Independent Institution -0.12 -0.11 0.17 -0.10 0.30 0.31 -0.23 -0.25
(-0.58) (-0.58) (0.66) (-0.37) (0.54) (0.58) (-1.02) (-1.33) HQ Country Norms -1.71 -1.99 -2.69 -2.75* 1.55 0.62 -1.22 -1.52 (-0.67) (-1.52) (-1.53) (-1.75) (0.51) (0.19) (-0.72) (-1.17)
Respondent Position FE Yes Yes Yes Yes Yes Yes Yes Yes Distribution Channel FE Yes Yes Yes Yes Yes Yes Yes Yes
Obs. 305 292 309 292 305 292 309 292 Pseudo R2 0.10 0.12 0.08 0.06 0.08 0.11 0.09 0.09
Internet Appendix: Survey Instrument
Survey on Climate Risk
We are a team of professors from the University of Geneva, the Swiss Finance Institute,
the University of Texas at Austin, and Frankfurt School of Finance & Management.
This survey seeks a better understanding of whether and how institutional investors
incorporate climate risk when making investment decisions. The survey will take about 10
minutes.
You can use this survey questionnaire or take the survey online at: [LINK]
We take the confidentiality of your responses very seriously. We will not share your
responses with anyone, nor will individual firms or respondents be identified. Only aggregate
data will be made public. We will not link the survey responses to any other data.
Thank you for participating in this survey. If you have any questions, please contact us.
Philipp Krueger, Ph.D. ([email protected])
Zacharias Sautner, Ph.D. ([email protected])
Laura T. Starks, Ph.D. ([email protected])
GENERAL INFORMATION
G1: How is the institution at which you work best described? □ Public pension fund □ Private pension fund
□ Insurance company □ Hedge fund
□ Mutual fund management company □ Private equity fund
□ Asset manager (for pension funds, endowments, etc.) □ Endowment, charity
□ Sovereign wealth fund □ Bank
□ Other (please specify): ____________________________
G2: What is the typical holding period for investments in your portfolio, on average? □ Short (less than 6 months)
□ Medium (6 months to 2 years)
□ Long (2 years to 5 years)
□ Very long (more than 5 years)
G3: What percentage of your portfolio is invested in fixed income versus equity securities? ___ % in fixed income
___ % in equities
G4: What percentage of your portfolio is invested actively versus passively? ___ % in active investments
___ % in passive investments
G5: What percentage of your portfolio incorporates Environmental, Social and Governance (ESG) issues? ____ %
G6: What is the total size of assets under management for your institution?
□ Less than $1 billion □ Between $1 billion and $20 billion
□ Between $20 billion and $50 billion □ Between $50 billion and $100 billion
□ More than $100 billion
G7: In which country are your institution’s headquarters based? ___________________________
G8: What is your position?
□ Fund/Portfolio Manager □ Chief Executive Officer
□ Investment Analyst/Strategist □ Executive/Managing Director
□ Chief Investment Officer □ ESG/Responsible Investment Specialist
□ CFO/COO/Chairman/Other Executive □ Other (please explain): _____________________
PART A: IMPORTANCE OF CLIMATE RISK
A1: Please rank the following six risks when making investments in portfolio firms from 1 to 6, where 1 is the most important to you and 6 the least important.
Financial risk (earnings, leverage, payout policy, etc.)
Operating risk (changes in demand, input costs, etc.)
Governance risk (board structure, executive pay, etc.)
Social risk (labor standards, human rights, etc.)
Climate risk
Other environmental risk (pollution, recycling, etc.)
A2: We have divided climate risk into regulatory risks (changes in regulation), physical risks (changes in the physical climate), and technological risks (climate-related technological disruption). Please rate the financial materiality of these risks.
Not at all
important
Slightly
important
Important Fairly
important
Very
important
Regulatory risks □ □ □ □ □
Physical risks □ □ □ □ □
Technological risks □ □ □ □ □
A3 to A5: [NOT COVERED IN THIS PAPER]
PART B: DISCLOSURE ON CLIMATE RISK
B1: How important do you consider reporting by portfolio firms on climate risk compared to reporting on financial information?
Much less
important
Less
important
Equally
important
More
important
Much more
important
□ □ □ □ □
B2: Do you disclose (or plan to disclose) the overall carbon footprint of your portfolio? □ No □ Yes □ Do not know
B3: To what extent do you agree with the following statements regarding climate-risk disclosure by portfolio firms? Strongly
agree
Agree Neither
agree
nor
disagree
Disagree Strongly
disagre
e
Investors should demand that portfolio firms disclose their exposure to climate risk
□ □ □ □ □
Firm-level quantitative information on climate risk is not sufficiently precise
□ □ □ □ □
Management discussions on climate risk are not sufficiently precise
□ □ □ □ □
Standardized and mandatory reporting on climate risk is necessary
□ □ □ □ □
Mandatory disclosure forms are not sufficiently informative regarding climate risk
□ □ □ □ □
There should be more standardization across markets in climate-related financial disclosure
□ □ □ □ □
Standardized disclosure tools and guidelines are currently not available
□ □ □ □ □
PART C: CLIMATE RISK MANAGEMENT & ENGAGEMENT
[THIS SECTION IS NOT COVERED IN THIS PAPER]
PART D: PRICING OF CLIMATE RISK
D1: To what extent do equity valuations of firms in different industries reflect the risks and opportunities related to climate change?
Industry
Valuations
much
too high
Valuations
somewhat
too high
Valuations
more or less
correct
Valuations
somewhat
too low
Valuations
much
too low
Oil □ □ □ □ □
Natural gas □ □ □ □ □
Renewable energy □ □ □ □ □
Nuclear energy □ □ □ □ □
Electric utilities □ □ □ □ □
Gas utilities □ □ □ □ □
Water utilities □ □ □ □ □
Coal mining □ □ □ □ □
Raw materials (excluding coal) □ □ □ □ □
Infrastructure □ □ □ □ □
Chemicals □ □ □ □ □
Automotive (traditional) □ □ □ □ □
Automotive (electric) □ □ □ □ □
Battery producers □ □ □ □ □
Construction □ □ □ □ □
Banking □ □ □ □ □
Insurance □ □ □ □ □
Agriculture □ □ □ □ □
Forestry and paper □ □ □ □ □
Information Technology □ □ □ □ □
Telecommunications □ □ □ □ □
Transportation □ □ □ □ □
Coastal real estate □ □ □ □ □
D2 to D4: [NOT COVERED IN THIS PAPER]
PART E: ADDITIONAL INFORMATION
E1: The Paris Climate Accord aims to keep the global temperature rise “well below 2 degrees Celsius” above pre-industrial levels by the end of this century. What are your expectations for the global temperature rise by the end of this century?
Increase in global temperature by:
Do not
know
None Up to 1
degree
Up to 2
degrees
Up to 3
degrees
More than
3 degrees
□ □ □ □ □ □
E2 to E4: [NOT COVERED IN THIS PAPER] E5: Do you engage (or plan to engage) portfolio companies to report according to the recommendations of the Task Force on Climate related Financial Disclosures (TCFD)?
□ No □ Yes □ Do not know